Indeed, retrograde optogenetics studies are rare. To help you, maybe look into AAV viruses of serotype 6 (AAV6), the only AAV serotype that is transported almost-exclusively retrogradely, in contrast to all other serotypes that are transported mostly anterogradely.

http://www.ncbi.nlm.nih.gov/pubmed/22418061 and http://www.ncbi.nlm.nih.gov/pubmed/24067867.

It seems the ChR2-coding construct is available from UNC. Do you want to target a genetically-identified population (ie. Cre-dependant)? Then you would need a FLEX or DIO version. I don't know if they exist.

Also, the Rabies virus is a good retrograde tool. If you were to inject a delta-G (=does't jump synapses) and non-pseudotyped (=non-EnvA so doesn't need a primo-sensitisation through a genetic construct or other virus) variant, you would achieve non-selective retrograde labeling from a given location. I currently use the EnvA chR2, but it must exist in non-EnvA, ChR2-coding version as well.

That's all I have for now. 

Best,

Thanks Julien, this is helpful :)

With regard to Rabies, I understood that it can cause major problems in the neurons it infects (comments on this would be appreciated), so I would rather use the AAV if possible.

Some other relevant publications included. You could drive cre expression from the target and then use a second cre inducible virus for ChR2 expression.

HSV: http://www.ncbi.nlm.nih.gov/pubmed/23636333

CAV2: http://www.ncbi.nlm.nih.gov/pubmed/16723393

PRV: http://www.ncbi.nlm.nih.gov/pubmed/25232307

Thanks for your input, Magor; this is interesting stuff indeed. However it seems simpler (at least for my purposes) to start with AAV6 and see how well it works.

In fact, re-reading the Zador group paper it is not clear to me what the advantage of their two-stage transfection technique (HSV-1 + AAV-2) when compared to a single step of injecting AAV6 (e.g., into the striatum in their case). Any comments on this?

When I was at UCSD we had a few people in our collaborating labs trying several different options with channelrhodopsin and retrograde labeling. We did not have HSV so it is one option we did not look into that much. Generally speaking, many retrograde labeling of AAV from the literatures seem to be hard to reproduce. We were told by the lab that tried this that purity/purification method and titer matter a lot - more than the serotype. Specific purification process can give great retrograde labeling. We did not have too much success with AAV overall, and the anterograde labeling is so dominating that retrograde labeling is irrelevant. Rabies virus (delta G) gives sparse labeling even at high titer. Great if you want to do anatomical studies with single cell labeled retrogradely, functionally limited. Toxicity is an issue - 3 days window for functional study. We end up with lentivirus with rabies glycoprotein - gives the highest number of cells per infection and easy to make. The idea of combining Cre and flexed-retrograde virus is potential to amplify the typically mediocre ChR2 expression. However, the execution of this approach is difficult, one low probability x another low probability = sparse labeling. Good luck with your experiment.

John, thanks for your input. Regarding your disapointing experience with AAV6, could you specify in which animals it was tried and in what region(s)?

Also, what promoter was used (if any)? It seems that people are getting good expression levels with CAG

A useful article comparing AAV serotypes from the Rumpel group (they find some retrograde labeling with AAV5): "Analysis of Transduction Efficiency, Tropism and Axonal Transport of AAV Serotypes 1, 2, 5, 6, 8 and 9 in the Mouse Brain"

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0076310

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0076310

For the most part retrograde labeling is not very reliable and may be brain region dependent. AAV9 has clear retrograde properties, but also anterograde and it's hard to predict where and how much retrograde is going on. AAV6 and 6.2 appear to be good choices as well, as mentioned by someone earlier.

A system that works quite well, primarily by Deisseroth but also others is a standard AAV expressing WGA-Cre. It's quite effective and we have used it for a variety of purposes. Here is one paper we published fairly recently with it (last figure)...

http://journal.frontiersin.org/article/10.3389/fnbeh.2015.00152/abstract

Thanks Caroline, very interesting stuff.

As for me, so far I have to report that, as John suggested, we have not been able to show retrograde transfection with AAV6s (2 different promoters and suppliers) when injected into the brainstem. My next step would be to try HSV, but it's always good to know what other options exist out there. BTW, with regard to this new paper by the Deisseroth team, has anyone else used CAV for such a purpose, or could comment on what might be it's advantages over other techniques?

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature15389.html

Hmm, just talked with collaborator and AAV6 is working beautifully for retrograde transport (Cre-AAV6 in a floxed GFP transgenic mouse). Something hinky may be going on with your source. I make my own, but many collaborators have complained recently about mixups with the primary core facilities. 

Somewhere I've read that 6.2 is even better but I can't find the link.

Since it's 2 different sources and the expression itself (at the site of injection, and in axons at sites to which the transfected neurons project) is very nice for both of them, I doubt the problem is the source. The difference may be in the species, the use of Cre, and/or brain areas...

Could I ask where did you get AAV6-DIO-ChR2? I looked at UNC, they do not have AAV6 type.

Dear Mengya,

At the time I ordered these vectors from UNC and from Stanford. It's possible that they are not sold anymore as a stock vector, although you could ask them directly. I'm assuming not so many people are interested in using AAV6 for retrograde gene delivery now that AAVretro is available (I haven't tried it yet, but some colleagues of mine have, and they say the results are quite good).

The original paper by the group that developed AAVretro at Janelia farm could be found here: Article A Designer AAV Variant Permits Efficient Retrograde Access t...

You could also find a recent comparison between these AAV pseudotypes by an independent group in this this paper:

Article Spinal nociceptive circuit analysis with recombinant adeno-a...

The AAV2/retro has much less "anterograde" expression compared to AAV2/6 or AAV2/9, which tend to have the most retrograde properties in the brain. This lack of anterograde could be helpful, or not, depends on the study design. Sometimes it's nice to have a marker of where the injection went. Other times it just confuses the data.

I have been seeing recent reports at meetings of folks using AAV2/5 for retrograde but find this to be very odd. My impression is that if you blow out a brain region with AAV2/5 (i.e. infuse a lot of virus) you can push the retrograde but this will likely only occur in a small subpopulation of neurons projection to the injection site. There could possibly be differences in packaging could promote retrograde properties of AAV2/5, but most people that I know using this serotype do not report retrograde transport.

Dear Colleague,

Retrograde transfection, a method for targeting gene expression in neuronal populations that project to specific areas, has been employed in various studies to express opsins in upstream neurons. This technique is particularly valuable in the field of optogenetics, where the expression of light-sensitive opsins in defined neuronal pathways is crucial for manipulating neuronal activity with spatial and temporal precision.

Several studies have successfully utilized retrograde viral vectors, such as adeno-associated viruses (AAVs) or lentiviruses engineered to preferentially infect projecting neurons, to express opsins in upstream neuronal populations. These retrograde vectors are injected into the target area, from which they are taken up by the axon terminals and transported back to the cell bodies of projecting neurons.

For instance:

• AAV-mediated Retrograde Transfection: Studies have used AAV vectors engineered with retrograde transport capabilities to express opsins like Channelrhodopsin-2 (ChR2) or Halorhodopsin in specific neuronal populations. By injecting these vectors into the brain region of interest, researchers have been able to modulate the activity of projecting neurons originating from distant areas.
• Use of Rabies Virus for Retrograde Targeting: Modified rabies viruses have been employed for retrograde transfection due to their natural ability to infect neurons trans-synaptically. This approach has facilitated the selective expression of opsins in upstream neurons, enabling the study of circuit-specific functions.

When seeking out specific studies, I recommend searching academic databases such as PubMed, Google Scholar, or the Web of Science for terms like "retrograde transfection opsins," "retrograde AAV optogenetics," or "retrograde viral vector expression in neurons." This search should yield articles and reviews detailing the use of retrograde transfection techniques to express opsins in upstream neurons for optogenetic manipulations.

Given the rapid advancements in viral vector technologies and optogenetics, reviewing recent publications and reviews can provide an up-to-date overview of the strategies, successes, and limitations of retrograde transfection methods. Additionally, considering the specific requirements of your experimental design, consulting with recent protocols and methodology papers will be invaluable in applying these techniques to your research.

Yours sincerely,

This protocol list might provide further insights to address this issue.